Bristol 5G city testbed with 5G-XHaul extensions

www.5g-xhaul-project.eu

INTRODUCING THE 5G-PPP 5G-XHAUL PROJECT Anna Tzanakaki (University of Bristol, NKUA)

• Huawei Technologies

• TU Dresden

• Telefónica I+D

• TES Electronic

Solutions

• University of Bristol

• University of Thessaly

1. CONSORTIUM OVERVIEW

2

• IHP GmbH (Coordinator)

• ADVA Optical Networking

• Airrays GmbH

• Blu Wireless Technology

• COSMOTE

• Fundació i2CAT

• Universities (3x), Research Institutes (2x), SMEs (2x), Operators (2x),

Industry partners (3x)

• 3 years duration

• Started: 01/07/2015

• 7.2 million euro EU funding

5G-XHAUL PHYSICAL INFRASTRUCTURE

The 5G-XHaul data plane considers an integrated optical and wireless network infrastructure for transport and access.

The wireless domain comprises small cells complemented by macro cells.

Fronthaul and backhauli can be supported through mmWave and Sub-6 wireless technologies or using a hybrid optical network platform combining both passive and active optical technologies.

3

VM

vBBU2

Data Centers

Small Cells

RU

RU

RU

RU

eNB

RU

RURU

RU

EPCS GW/GSN

vBBU

vBBUvBBU

PDN-GW

vBBU1

Backhaul

Fronthaul

vBBU

vBBU1 vBBU2

VM

eNB

Macro Cell

Internet

60GHz /Sub-6 links for FH/BH

FHBH

HeNBFemto Cells

HeNBHeNB GW

HeNB

WiFiWLAN GW

TSON

WDM-PON

5G-Xhaul Wireless Backhaul/Fronthaul

HeNB: Home eNodeB

Wireless Access

VM: Virtual Machine

GW: Gateway

vBBU: Virtual Base Band Unit

RU: Remote Unit

PDN-GW: Packet Data Network GateWay

EPC: Evolved Packet Core

S GW: Serving GateWay

5G-XHAUL INPUT TO THE 5G PPP VIEW ON 5G ARCHITECTURE

• 5G PPP View on 5G Architecture – (White

Paper), https://5g-ppp.eu/white-papers/

• 5G PPP View on 5G Architecture - Section

5 - Physical architecture, V. Jungnickel,

Fraunhofer HHI, WORKSHOP 1:

International Workshop on 5G

Architecture, EuCNC 2016

MP2MP

MP2MP

Elastic Frame-based WDM Metro

FlexGridROADM

FlexGridROADM

FlexGridROADM

FlexGridROADM

MP2MP

MP2MP

5G-XHAUL OVERARCHING LAYERED ARCHITECTURE

5

A. Tzanakaki et al., "5G infrastructures supporting end-user and operational services: The 5G-XHaul architectural perspective," 2016 IEEE International Conference on Communications Workshops (ICC), Kuala Lumpur, Malaysia, 2016, pp. 57-62

End Point A

VM

End Point B

vBBU (upper layers)

Virtual Wireless

forwarding

Virtual Optical

Virtual BH

Virtual Wireless

forwarding

Virtual Optical

Virtual FH

RRH

RRHRRH

VM

vBBUInternet

Fronthaul

Optical Transport Data Centers

Backhaul

Data Centers

vBBU

Wireless Access/Transport

VM

mmWave Network

Controller

LTE Network Controller

Compute Controller

WDM PON Controller

TSON Controller

Compute Controller

Virtualization Virtualization Virtualization Virtualization

Virtual Network

Physical Network

ProcessingVirtual

Network

Physical Network

Processing

SDN Controller VNF

SDN Controller

VNF

EM EMElement Management

(EM) Element Management (EM)

DriversMa

na

ged

PH

Y I

nf

Inf

Man

age

me

nt

RU

RF to Baseband

Cycle Prefix & FFT

Resource demapping

Receive processing

Decoding

MAC

(1)

(2)

(3)

(4)

(5)

Clo

ud

RA

NT

rad

itio

na

l R

AN

high network bandwidth Increased BBU sharing

low network bandwidth Limited BBU sharing

Co

ntr

ol

Ma

na

gem

en

t &

Se

rvic

e

Orc

he

stra

tio

n

vBBU (lower layers)

Development focus

ARCHITECTURE PERFORMANCE EVALUATION

6

mmWave links Optical fiber TSON nodes Large scale DC

Bristol 5G city network topology with mmWave backhauling

Snapshot of spatial traffic load

Mullti-objevctive optimisation model aims to identify the optimal resources and policies that can support the required services in terms of both topology and resources. Optimal FH and BH service provisioning, with the overall objective to maximise the energy efficiency of the infrastructure and minimize end-to-end service delays.

A. Tzanakaki et all, “Wireless-Optical Network Convergence: Enabling the 5G Architecture to Support Operational and End-User Services“ IEEE Comms Magazine, August 2017

NUMERICAL RESULTS: ENERGY-DELAY

• Figs a) and d): average traffic per BS and spatial traffic distribution for the wireless access domain

• The C-RAN approach offers significant energy savings (60-75%) (Fig. b)

• Overloading of network resources to support FH, the C-RAN case increases the end-to-end service delay in the BH (Fig. c), which remains below 20ms for a 100 Mbps flow request

• The BH service delay for C-RAN vBBU is lower compared to the delay for the C-RAN fixed BBU case

0

5

10

0

5

10

20

40

60

80

Km

Ave

rag

e d

ata

ra

te (

Mb

ps) 0 20 40 60 80

X (km)Y (km)d)

DATA-PLANE: WIRELESS

mmWave (60GHz) Front End design Antenna & BFIC

mmWave Base Band design

MIMO/Beam alignment and tracking/P2MP Channel modelling

Synchronization in wireless backhaul: 1588v2, ToF based

Functional splits for 5G-RANs (NGFI):

Impact on transport requirements Specific development for Massive MIMO

Self-backhauling: Joint access and backhaul

5G-XHaul mmWave

BFIC

5G-XHaul mmWave

nodes

Massive MIMO array

supporting 5G-XHaul

functional split

8

DATA-PLANE: OPTICAL

Hybrid passive/active optical network solution supporting joint FH & BH

Active: Time Shared Optical Networks (TSON)

Elastic BW allocation (time slices)

Extensions for elastic grid

Native mapping of Ethernet and CPRI

Synchronization

Passive: flexible WDM-PON

40λs, 10-25 Gbps/λ, 20-40 Km

Color-less ONUs (out-of-band mgmt)

Switch off ONUs for energy saving

Flexible assignment BBU-RRH

TSON FPGA

implementation

OLT

Tx Array

...

1

RN

L C

n

DE

MU

X

Rx Array

...

Cyclic A

WG

MU

X

ONU n

T-LD

Rx

L C

ONU 1

T-LD

Rx

L C

2

1

n

2

BBU

BBU

…

Ethernet

switch

...

... Cro

ss-c

on

nect

OLT

5G-XHaul WDM-PON architecture 9

Testbed configuration of TSON and WDM-PON

Integration using BIO dark fiber

DATA-PLANE: WDM-PON & TSON INTEGRATION

10

TSON

Node 2 SFP+

SFP+

Cro

ss-c

onnect

SMFDark fibre

13

10

nm

SMFTSON

Node 1

SFP+

Tx

RxTx

Rx

L-b

an

d

Tx

Rx

Tx

Rx

Tx

Rx

SMFRx

Tx

OLT

ONU

DEM

UX

MU

XD

EM

UX

MU

X

SMFDark fibre

Eth

ern

etA

nal

yser

Ch.1

Tx

Rx

Rx

Tx

Rx

Tx

Ch.2Tx

RxTr

an

spo

nd

erWDM-PON link

RN

TSON Network

C-band

Bristo

lIsO

pen (B

IO)

Dark

fibre

13

10

nm

L-b

an

d

C-b

an

d u

pst

rea

m

L-b

an

d d

ow

nst

rea

m

DC

F

EDFA

VCSEL SFP+

TSON Node 1&2 ONU OLT side

Downstream Latency

DATA-PLANE: MASSIVE MIMO FH OVER WDM-PON

11

CPRI over WDM ADVA

Baseband unit (I/Q samples)

AIR

RF cable

RRH

AIR

UE receiver

TUD

Colored

DWDM

Tra

nsp

on

der

λ-agnostic

DWDM

ONU

…

ONU

…

CPRI

Wireless & Mobile

Net.

Wi-Fi 802.11ac, LTE, mmWave,

Massive MIMO, 60GHz backhaul

RF Mesh Network

8 Fiber-connected lampposts

with 1,500 photocells and

any-sensor hosting capability

Computing

Infrastructure

HPC facility, commodity

compute/storage, private cloud

and edge mobile computing

Optical Network

144-fiber core network

connecting 4 active nodes,

full optical switching, flexi optical

CITY-TRIALS: BIO INFRASTRUCTURE

OLT

ON

U

TSO

N

TSO

N

mmWav

e

Sub6

Hotspo

t

Massive

MIMO

SAMPLE OF PLANNED DEMONSTRATION IN BRISTOL (JUNE’18)

13

BBU

COMPUTE

TSO

N

FRONTHAUL (CPRI)

BACKHAUL (ETH)

Thanks for

your attention!

Questions?

www.5g-xhaul-project.eu